Every strategy game I grew up with taught me the same quiet lesson, and it took years to notice it.
You scout. You read the composition. You pick the engagement, you get the angle, you commit at the right moment — and then a number nobody can see decides that your marksman misses three times in a row and the fight goes the other way.
The genre calls this tension. I think it is closer to noise.
What randomness actually teaches
A random roll in combat is not neutral. It is a lesson, repeated a few hundred times an hour, and the lesson is this: volume beats precision.
If any individual engagement can be overturned by chance, the rational response is to reduce the influence of any individual engagement. Send more. Trade more. Stop reading the board so carefully, because careful reading is expensive and the dice do not care how long you spent on it.
That is a coherent strategy. It is just not the strategy I wanted the game to be about.
The rule
In Longbarrow, no die is rolled in combat.
No accuracy percentages. No critical hits. No hidden variance smoothing out over a long enough fight. Damage, range, speed, cover and vision resolve identically every time. Set up the same engagement twice and it plays out the same way twice.
This sounds like it removes drama. It does the opposite — it relocates it. The tension moves off the roll and onto the decision, where you can actually do something about it.
What determinism buys
Every loss becomes information. When chance cannot be blamed, something about your position, timing, spacing or composition was worse than theirs. That is uncomfortable, and it is the most useful feedback a strategy game can give you.
Reports can be honest. The post-battle summary does not need to estimate a confidence or hedge about variance. It can tell you what happened, because what happened is the only thing that could have happened.
Terrain becomes an argument. High ground extends what you can see rather than improving what you hit — a real informational advantage instead of a hidden multiplier. Bushes conceal. Cover is soft or hard and behaves differently. Trees and cliffs come down. None of that has to compete with a random number for relevance.
Small forces can win. Clumping suffers diminishing returns, so a smaller force in the right formation, on the right ground, at the right moment, genuinely beats a larger one that arrived badly. That outcome is earned rather than rolled, which means it is repeatable — and a repeatable upset is a skill, not a story.
The part that is hard
Determinism is unforgiving to the designer. Every unit is a solved problem waiting to happen: if the numbers resolve identically every time, then somewhere in the matchup table there is an optimal answer, and players will find it faster than I can patch it.
Randomness papers over that. It buys designers slack by making the optimum fuzzy.
Taking the dice away means the balance has to come from somewhere real — from terrain, timing, spacing, vision, and the cost of being in the wrong place. That is more work. It is also the actual game.
I would rather build the harder version and have the better decision win.